DOCSLIB.ORG

Distributed Algorithms for Covering, Packing and Maximum Weighted Matching

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Distributed Algorithms for Covering, Packing and Maximum Weighted Matching Noname manuscript No. (will be inserted by the editor) Distributed Algorithms for Covering, Packing and Maximum Weighted Matching Christos Koufogiannakis · Neal E. Young Received: date / Accepted: date Abstract This paper gives poly-logarithmic-round, dis- 1 Background and results tributed δ-approximation algorithms for covering prob- lems with submodular cost and monotone covering con- Many distributed systems are composed of components straints (Submodular-cost Covering). The approx- that can only communicate locally, yet need to achieve a imation ratio δ is the maximum number of variables in global (system-wide) goal involving many components. any constraint. Special cases include Covering Mixed The general possibilities and limitations of such sys- Integer Linear Programs (CMIP), and Weighted tems are widely studied [42,49,54,38,39,12]. It is of Vertex Cover (with δ = 2). specific interest to see which fundamental combinato- Via duality, the paper also gives poly-logarithmic- rial optimization problems admit efficient distributed round, distributed δ-approximation algorithms for Frac- algorithms that achieve approximation guarantees that tional Packing linear programs (where δ is the max- are as good as those of the best centralized algorithms. imum number of constraints in which any variable oc- Research in this spirit includes works on Set Cover curs), and for Max Weighted c-Matching in hy- (Dominating Set) [26,40,41], capacitated domi- pergraphs (where δ is the maximum size of any of the nating set [37], Capacitated Vertex Cover [16, hyperedges; for graphs δ = 2). 17], and many other problems. This paper presents dis- The paper also gives parallel (RNC) 2-approximation tributed approximation algorithms for some fundamen- algorithms for CMIP with two variables per constraint tal covering and packing problems. and Weighted Vertex Cover. The algorithms use the standard synchronous com- The algorithms are randomized. All of the approx- munication model: in each round, nodes can exchange imation ratios exactly match those of comparable cen- a constant number of messages with neighbors and per- tralized algorithms.1 form local computation [54]. There is no restriction on message size or local computation. The algorithms are Keywords Approximation algorithms · Integer linear efficient | they finish in a number of rounds that is programming · Packing and covering · Vertex cover · poly-logarithmic in the network size [42]. Matching C. Koufogiannakis 1.1 Covering Problems Department of Computer Science and Engineering University of California, Riverside Consider optimization problems of the following form: given a non-decreasing, continuous, and submodular2 N. E. Young cost function c : IRn ! IR , and a set of constraints C Department of Computer Science and Engineering + + 3 University of California, Riverside where each constraint S 2 C is closed upwards , 1 n Preliminary versions appeared in [34,35]. Work by the find x 2 IR+ minimizing c(x) s.t. (8S 2 C) x 2 S. first author was partially supported by the Greek State Scholarship Foundation (IKY). Work by the second author 2 Formally, c(x) + c(y) ≥ c(x ^ y) + c(x _ y) where ^ and _ was partially supported by NSF awards CNS-0626912, CCF- are component-wise minimum and maximum, respectively. 0729071. 3 If x 2 S and y ≥ x, then y 2 S. 2 Christos Koufogiannakis, Neal E. Young Submodular-cost Covering includes all problems vertex cover) can be computed in an expected O(log n) of this form [36]. The (approximation) parameter δ is rounds via the algorithm of Israeli and Itai [25]. the maximum number of elements of x on which any Maximal Matching is also in NC [9,47,30] and in constraint x 2 S depends. RNC | parallel poly-log time with polynomially many n In the above formulation, x ranges over IR+, but, be- randomized processors [25] | hence so is 2-approximat- cause the constraints can be non-convex, one can handle ing Unweighted Vertex Cover. arbitrarily restricted variable domains by incorporating For Weighted Vertex Cover, previous to this them into the constraints [36]. For example, Weighted work, no efficient distributed 2-approximation algorithm P Vertex Cover is equivalent to minimize v cvxv sub- was known. Similarly, no parallel NC or RNC 2-ap- n ject to x 2 IR+ and proximation algorithm was known. in 1994, Khuller, Vishkin and Young gave 2(1 + ")-approximation algo- xu ≥ 1 or xw ≥ 1 (8(u; w) 2 E): rithms: a distributed algorithm taking O(log n log 1=") (Given any 2-approximate solution x to this formula- rounds and a parallel algorithm, in NC for any fixed tion | which allows x 2 IR | rounding each x down u + u " [31].4 Given integer vertex weights, their result gives to its floor gives a 2-approximate integer solution.) distributed 2-approximation in O(log n log nC^)) rounds, Submodular-cost Covering includes the follow- where C^ is the average vertex weight. The required ing problems as special cases: number of rounds is reduced to (expected) O(log nC^) CIP, covering integer programs with variable upper by Grandoni, K¨onemannand Panconesi [18,15]. m×n m n bounds: given A 2 IR+ , w 2 IR+ , u 2 IR+, As noted in [38], neither of these algorithms is effi- n cient (taking a number of rounds that is polylogarith- minimize c · x subject to x 2 ZZ+, Ax ≥ w, and x ≤ u. mic in the number of vertices). CMIP, covering mixed integer linear programs: CIP Kuhn, Moscibroda and Wattenhofer describe dis- with both integer and fractional variables. tributed approximation algorithms for fractional cov- Facility Location, Weighted Set Cover (that is, ering and packing linear programs [39]. They show an CIP with Aij 2 f0; 1g; wi = 1), Weighted Ver- O(1)-approximation with high probability (w.h.p.) in tex Cover (that is, Weighted Set Cover with O(log m) rounds (m is the number of covering con- δ = 2), and probabilistic (two-stage stochastic) vari- straints). The approximation ratio is greater than 2 for ants of these problems. Fractional Weighted Vertex Cover. For (inte- In the centralized (non-distributed) setting there are ger) Weighted Vertex Cover and Weighted Set two well-studied classes of polynomial-time approxima- Cover (where each Aij 2 f0; 1g) combining their al- tion algorithms for covering problems: gorithms with randomized rounding gives O(log ∆)-ap- proximate integer solutions in O(log n) rounds, where (i) O(log ∆)-approximation algorithms where ∆ is the ∆ is the maximum number of constraints in which any maximum number of constraints in which any vari- variable occurs. able occurs (e.g. [27,46,10,56,57,32]), and Distributed lower bounds. The best lower bounds (ii) O(δ)-approximation algorithms where δ is the max- known for Weighted Vertex Cover are by Kuhn, imum number of variables in any constraint (e.g. [4, Moscibroda and Wattenhofer: to achieve even a poly- 22,21,5,19,7,8,55,2,36]), including most famously logarithmic approximation ratio requires in the worst 2-approximation algorithms for Weighted Ver- case Ω(plog n= log log n) rounds. In graphs of constant tex Cover. degree ∆, Ω(log ∆/ log log ∆) rounds are required [38]. The algorithms here match those in class (ii). New results for covering problems. This paper Related work. For Unweighted Vertex Cover, gives the following results for covering problems. it is well known that a 2-approximate solution can be { Section 2 describes the first efficient distributed 2- found by computing any maximal matching, then tak- approximation algorithm for Weighted Vertex ing the cover to contain the endpoints of the edges in the Cover. The algorithm runs in O(log n) rounds in matching. A maximal matching (and hence a 2-approx- expectation and with high probability. imate vertex cover) can be computed deterministically in O(log4 n) rounds using the algorithm of Ha´n´ckowiak, { Section 3, generalizing the above result, describes Karonski and Panconesi [20] or in O(∆+log∗ n) rounds the first efficient distributed 2-approximation algo- using the algorithm of Panconesi and Rizzi [51], where 4 Their result extends to δ(1 + ")-approximating Weighted ∆ is the maximum vertex degree. Or, using randomiza- Set Cover: a distributed algorithm taking O(δ log n log 1=") tion, a maximal matching (and hence a 2-approximate rounds, and a parallel algorithm (in NC for fixed " and δ). Distributed Algorithms for Covering, Packing and Maximum Weighted Matching 3 problem approx. ratio # rounds where when 2 O(log n) (random) [25] 1986 Unweighted Vertex Cover 2 O(log4 n) [20] 2001 2 + "O(log "−1 log n) [31] 1994 Weighted Vertex Cover 2 O(log n log nC^) [31] 1994 2 O(log nC^) (random) [18,15] 2005 2 O(log n) (random) here Weighted Set Cover O(log ∆) O(log m) (random) [39] 2006 δ O(log m) (random) here CMIP (with δ = 2) 2 O(log m) (random) here CMIP δ O(log2 m) (random) here Table 1 Comparison of distributed algorithms for covering problems. δ is the maximum number of variables in any constraint. ∆ is the maximum number of constraints in which any variable occurs. rithm for CMIP (covering mixed integer linear pro- of generality each vertex capacity cj is also an integer.) grams with variable upper bounds) restricted to in- An instance is defined by a given hypergraph H(V; E) jV j stances where each constraint has at most two vari- and cost vector c 2 ZZ+ ; a solution is given by a vector ables (δ = 2). The algorithm runs in O(log m) rounds jEj P y 2 ZZ+ maximizing e2E weye and meeting all the in expectation and with high probability, where m P vertex capacity constraints e2E(u) ye ≤ cu (8u 2 V ), is the number of constraints.
Load more

Recommended publications
  • On Vertex Independence Number of Uniform Hypergraphs
    Acta Univ. Sapientiae, Informatica, 6, 1 (2014) 132{158 On vertex independence number of uniform hypergraphs Tariq A. CHISHTI Guofei ZHOU University of Kashmir Nanjing University Department of Mathematics Department of Mathematics Srinagar, India Nanjing, China email: [email protected] email: [email protected] Shariefuddin PIRZADA Antal IVANYI´ University of Kashmir E¨otv¨osLor´andUniversity Department of Mathematics Faculty of Informatics Srinagar, India Budapest, Hungary email: email: [email protected] [email protected] Abstract. Let H be an r-uniform hypergraph with r ≥ 2 and let α(H) be its vertex independence number. In the paper bounds of α(H) are given for different uniform hypergraphs: if H has no isolated vertex, then in terms of the degrees, and for triangle-free linear H in terms of the order and average degree. 1 Introduction to independence in graphs Let n be a positive integer. A graph G on vertex set V = fv1; v2; : : : ; vng is a pair (V; E), where the edge set E is a subset of V × V. n is the order of G and jEj is the size of G. Computing Classification System 1998: G.2.2 Mathematics Subject Classification 2010: 05C30, 05C50 Key words and phrases: uniform hypergraph, independence number, lower bound DOI:10.2478/ausi-2014-0022 132 On vertex independence number of uniform hypergraphs 133 Let v 2 V and N(v) be the neighborhood of v, namely, the set of vertices x so that there is an edge which contains both v and x. Let U be a subset of V, then the subgraph of G induced by U is defined as a graph on vertex set U and edge set EU = f(u; v)ju 2 u and v 2 Ug.
    [Show full text]
  • Qualitative Reasoning for Spatial Matching 1 [email protected] Qualitative Reasoning for Spatial Matching
    Qualitative Reasoning for Spatial Matching 1 [email protected] Qualitative Reasoning for Spatial Matching Diedrich Wolter Universit¨atBremen, Transregional Collaborative Research Center SFB/TR 8 Spatial Cognition Abstract. Matching two representations against one another is central to many applications, e.g., matching observation against models in ob- ject or scene recognition or matching a local perception against a global map in self-localization if a mobile robot. In both examples, spatial infor- mation structures are associated, i.e., the representation includes spatial arrangement information of individual parts, either in a relative or an absolute frame of reference. In this paper the utility of qualitative configuration information for matching is investigated. Qualitative configuration information expresses relative arrangement information. It is argued that such information is helpful for making side conditions of plausible matchings explicit, which can improve matching efficiency as well as quality of the matching. Keywords. Matching, Qualitative Spatial Information 1 Introduction Many applications involve the task of matching two representations against one another. For example scene recognition involves identifying a set of spatially distributed objects. Many objects bear rich information which can be exploited the recognition task, but it is information about the spatial arrangement of the involved objects that allows for important distinctions. Consider you are observing a scene that contains dishes, glasses, and sets of cutlery. These objects, when nicely arranged on a table, gives rise to the hypothesis that you are about to enjoy a nice meal, whereas piled up dishes, glasses, and clutter of cutlery suggests dish washing to be your destiny. This article investigates into matching tasks which are involved with match- ing spatially embedded representations, i.e.
    [Show full text]
  • Curriculum Vitae Prof. Dr. Anand Srivastav Personal Data Name Prof
    Curriculum Vitae Prof. Dr. Anand Srivastav Personal Data Name Prof. Dr. Anand Srivastav (Dipl. Math., Dipl. Phys.) Place/Date of Birth Agra (India), in 1958 Affiliation Institution Christian-Albrechts-Universität zu Kiel (CAU) Institute Department of Computer Science Address Christian-Albrechts-Platz 4, D-24118 Kiel, Germany Phone / Fax +49 431 880-7252 / -1725 E-Mail [email protected] Education 1996 Habilitation in Computer Science, Freie Universität zu Berlin, Germany 1988 Dr. rer. nat. (in Mathematics), Westfälische Wilhelms-Universität Münster 1988 Dipl. Phys. degree (Master of Physics), Westfälische Wilhelms-Universität Münster, Germany 1984 Dipl. Math. degree (Master of Mathematics), Westfälische Wilhelms- Universität Münster, Germany 1978 – 1984 Study of Physics and Mathematics at the Westfälische Wilhelms- Universität Münster, Germany 1965 – 1978 School Education in India and Germany Academic Appointments since 1997 Professor for Mathematics (Discrete Optimization), Christian-Albrechts- Universität zu Kiel 1995 – 1997 Assistent Professor at the Computer Science Department, Humboldt- Universität zu Berlin 1994 – 1995 Visiting Professor at the Computer Science Department, Freie Universität Berlin 1993 – 1994 Visiting Professor at IMA (University of Minnesota), Courant Institute for Mathematical Sciences (New York University) and Yale University, USA 1988 – 1993 Assistent Professor, Institute for Discrete Mathematics, Rheinische Friedrich- Wilhelms-Universität Bonn Important Functions since 2008 Co-Speaker of the Research
    [Show full text]
  • Matchings in Balanced Hypergraphs
    Matchings in balanced hypergraphs Von der Fakultat¨ fur¨ Mathematik, Informatik und Naturwissenschaften der RWTH Aachen University zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften genehmigte Dissertation vorgelegt von Diplom-Mathematiker Robert Berthold Scheidweiler aus Duren-Birkesdorf.¨ Berichter: Universitatsprofessor¨ Dr. Eberhard Triesch Universitatsprofessor¨ Dr. Arie M.C.A. Koster Tag der mundlichen¨ Prufung:¨ 19. April 2011 Diese Dissertation ist auf den Internetseiten der Hochschulbibliothek online verfugbar.¨ Danksagung In den letzten funf¨ Jahren habe ich am Lehrstuhl II fur¨ Mathematik der RWTH Aachen Uni- versity die vorliegende Dissertation verfasst. Einigen gebuhrt¨ fur¨ ihre Unterstutzung¨ und Hilfe wahrend¨ dieser Zeit besonderer Dank. Zuallererst mochte¨ ich mich bei dem Betreuer meiner Dissertation und meinem Chef, Eberhard Triesch, bedanken. Durch ihn habe ich das Thema dieser Arbeit erhalten, das mir sehr ans Herz gewachsen ist. Er hat mir bei meinen Forschungen immer mit Rat und Tat zur Seite gestanden und mich auch bei langer¨ andauern- den Durststrecken niemals unter Druck gesetzt. Seine positive Unterstutzung¨ und geduldige Hilfe haben mich motiviert, diese Arbeit zu vollenden. Weiterhin mochte¨ ich mich bei Arie Koster, meinem Zweitgutachter, bedanken. Mehrfach hat er im Verlauf meiner Promotion Anregungen gegeben, die dann in die Dissertation eingeflossen sind. Vor der endgultigen¨ Ab- gabe hat er durch seine Verbesserungsvorschlage,¨ fur¨ die ich sehr dankbar bin, zur jetzigen Form der Arbeit beigetragen. Danken mochte¨ ich außerdem Bert Randerath, der mir half, einige Startschwierigkeiten zu uberwinden,¨ als ich begann, die balancierten Hypergraphen zu erforschen. Hartmut Fuhr¨ hat sehr viel Zeit darauf verwendet, mir die harmonische Analysis naher¨ zu bringen. Seine Bemuhungen¨ haben meine Promotion weiter voran gebracht.
    [Show full text]
  • Learning from Networked Examples
    Learning from Networked Examples∗ Yuyi Wang ETH Zurich,¨ Switzerland Zheng-Chu Guo Zhejiang University, P.R. China Jan Ramon Inria, France & KULeuven, Belgium Editor: The correct dates will be entered by the editor Abstract Many machine learning algorithms are based on the assumption that training examples are drawn independently. However, this assumption does not hold anymore when learning from a networked sample because two or more training examples may share some common objects, and hence share the features of these shared objects. We show that the classic approach of ignoring this problem potentially can have a harmful effect on the accuracy of statistics, and then consider alternatives. One of these is to only use independent examples, discarding other information. However, this is clearly suboptimal. We analyze sample error bounds in this networked setting, providing signifi- cantly improved results. An important component of our approach is formed by efficient sample weighting schemes, which leads to novel concentration inequalities. Keywords: Networked Examples, Non-i.i.d. Sample, Concentration Inequality, Sample Error Bound, Generalization Error Bound 1. Introduction Recently, there has been an increasing interest in network-structured data, such as data in social net- works (Leinhardt, 2013), economic networks (Knoke, 2014), citation networks (Nykl et al., 2014) and chemical interaction networks (Szklarczyk et al., 2014). An important challenge is that data concerning related objects cannot be assumed to be independent. More precisely, a key assump- arXiv:1405.2600v4 [cs.AI] 3 Jun 2017 tion made by many approaches in the field of statistics and machine learning is that observations are drawn independently and identically (i.i.d.) from a fixed distribution.
    [Show full text]
  • Deterministic Distributed Edge-Coloring Via Hypergraph Maximal Matching
    Deterministic Distributed Edge-Coloring via Hypergraph Maximal Matching Manuela Fischer Mohsen Ghaffari Fabian Kuhn ETH Zurich ETH Zurich University of Freiburg manuela.fi[email protected] ghaff[email protected] [email protected] Abstract We present a deterministic distributed algorithm that computes a (2∆ 1)-edge-coloring, or even list-edge-coloring, in any n-node graph with maximum degree ∆, in O(log− 7 ∆ log n) rounds. This answers one of the long-standing open questions of distributed graph algorithms· from the late 1980s, which asked for a polylogarithmic-time algorithm. See, e.g., Open Problem 4 in the Distributed Graph Coloring book of Barenboim and Elkin. The previous best round complexities O(√log n) were 2 by Panconesi and Srinivasan [STOC’92] and O˜(√∆) + O(log∗ n) by Fraigniaud, Heinrich, and Kosowski [FOCS’16]. A corollary of our deterministic list-edge-coloring also improves the randomized complexity of (2∆ 1)-edge-coloring to poly(log log n) rounds. − The key technical ingredient is a deterministic distributed algorithm for hypergraph maximal matching, which we believe will be of interest beyond this result. In any hypergraph of rank r — where each hyperedge has at most r vertices — with n nodes and maximum degree ∆, this algorithm computes a maximal matching in O(r5 log6+log r ∆ log n) rounds. · This hypergraph matching algorithm and its extensions also lead to a number of other results. In particular, we obtain a polylogarithmic-time deterministic distributed maximal independent set (MIS) algorithm for graphs with bounded
    [Show full text]
  • Tur\'An Numbers for Forests of Paths in Hypergraphs
    TURAN´ NUMBERS FOR FORESTS OF PATHS IN HYPERGRAPHS NEAL BUSHAW AND NATHAN KETTLE Abstract. The Tur´an number of an r-uniform hypergraph H is the maximum number of edges in any r-graph on n vertices which does not contain H as a (r) subgraph. Let Pℓ denote the family of r-uniform loose paths on ℓ edges, F(k,l) (r) (r) denote the family of hypergraphs consisting of k disjoint paths from Pℓ , and Lℓ denote an r-uniform linear path on ℓ edges. We determine precisely exr(n; F(k,l)) (r) and exr(n; k · Lℓ ), as well as the Tur´an numbers for forests of paths of differing lengths (whether these paths are loose or linear) when n is appropriately large dependent on k,l,r, for r ≥ 3. Our results build on recent results of F¨uredi, Jiang, and Seiver who determined the extremal numbers for individual paths, and provide more hypergraphs whose Turan numbers are exactly determined. 1. Introduction and Background Extremal graph theory is that area of combinatorics which is concerned with find- ing the largest, smallest, or otherwise optimal structures with a given property. Often, the area is concerned with finding the largest (hyper)graph avoiding some subgraph. We build on earlier work of F¨uredi, Jiang, and Seiver [19], who deter- mined the extremal numbers when the forbidden hypergraph is a single linear path or a single loose path. In this paper, we determine precisely the exact Tur´an numbers when the forbidden hypergraph is a forest of loose paths, or a forest of linear paths; our main results appear in Section 2.
    [Show full text]
  • On Derandomizing Local Distributed Algorithms
    On Derandomizing Local Distributed Algorithms Mohsen Ghaffari David G. Harris Fabian Kuhn∗ ETH Zurich University of Maryland, College Park University of Freiburg [email protected] [email protected] [email protected] Abstract The gap between the known randomized and deterministic local distributed algorithms underlies ar- guably the most fundamental and central open question in distributed graph algorithms. In this paper, we develop a generic and clean recipe for derandomizing randomized LOCAL algorithms and transform- ing them into efficient deterministic LOCAL algorithm. We also exhibit how this simple recipe leads to significant improvements on a number of problems, in cases resolving known open problems. Two sample end-results are as follows: • An improved distributed hypergraph maximal matching algorithm, which improves on that of Fis- cher, Ghaffari, and Kuhn [FOCS’17], and leads to improved algorithms for edge-coloring, maxi- mum matching approximation, and low out-degree edge orientation. The first gives an improved algorithm for Open Problem 11.4 of the book of Barenboim and Elkin, and the last gives the first positive resolution of their Open Problem 11.10. • An improved distributed algorithm for the Lovasz´ Local Lemma, which gets closer to a conjec- ture of Chang and Pettie [FOCS’17], and moreover leads to improved distributed algorithms for problems such as defective coloring and k-SAT. ∗Supported by ERC Grant No. 336495 (ACDC). 1 Introduction and Related Work The gap between the best known deterministic and randomized local distributed algorithms constitutes a central open question in the area of distributed graph algorithms. For many of the classic problems (e.g., maximal independent set (MIS) and (∆ + 1)-vertex coloring), O(log n)-time randomized algorithms have been known since the pioneering work of Luby [Lub86] and Alon, Babai, and Itai [ABI86].
    [Show full text]
  • Matching Theory-A Sampler: from D&Es Kgnig to The
    Discrete Mathematics 100 (1992) 177-219 177 North-Holland Matching theory-a sampler: from D&es KGnig to the present Michael D. Plummer* Department of Mathematics, Vanderbilt University, Nashville, TN 37235, USA Received 20 March 1991 Revised 8 October 1991 Abstract Plummer, M.D., Matching theory-a sampler: from D&es Kijnig to the present, Discrete Mathematics 100 (1992) 177-219. 1. Introduction and terminology Let G be a finite undirected graph without loops or multiple lines. A set of lines M c E(G) is a matching if no two share a common endpoint. A matching is said to be perfect if it covers all points of G. One can take the point of view that a perfect matching is a special case of a more general concept in graph theory-an n-factor. An n-factor of graph G is a spanning subgraph which is regular of degree n; that is, each point has degree it. So a perfect matching is a l-factor. But we can generalize even more. Let f denote a function which assigns to each point u in V(G) a nonnegative integer f(v). Then an f-factor is any subgraph F of G in which deg F2r =f(v). The existence of these increasingly more general concepts for a fixed graph G are all instances of what have come to be called degree-constrained subgraph problems. The reader of these Proceedings may see in Mulder’s article [126] that these ideas were of great interest to Julius Petersen and in fact Petersen enjoyed some considerable success in his studies of such problems.
    [Show full text]
  • New Applications of the Nearest-Neighbor Chain Algorithm
    UC Irvine UC Irvine Electronic Theses and Dissertations Title New Applications of the Nearest-Neighbor Chain Algorithm Permalink https://escholarship.org/uc/item/6nt3g7d4 Author Mamano Grande, Nil Publication Date 2019 Peer reviewed|Thesis/dissertation eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA, IRVINE New Applications of the Nearest-Neighbor Chain Algorithm DISSERTATION submitted in partial satisfaction of the requirements for the degree of DOCTOR OF PHILOSOPHY in Computer Science by Nil Mamano Grande Dissertation Committee: Professor Michael T. Goodrich, Chair Professor David Eppstein Professor Sandy Irani 2019 c 2019 Nil Mamano Grande TABLE OF CONTENTS Page LIST OF FIGURES v LIST OF TABLES vii LIST OF ALGORITHMS viii ACKNOWLEDGMENTS ix CURRICULUM VITAE x ABSTRACT OF THE DISSERTATION xii 1 Introduction 1 1.1 Greedy algorithms . 3 1.2 Local greedy . 4 1.3 Global–local equivalence . 6 1.4 Background . 8 1.4.1 Hierarchical clustering . 8 1.4.2 Nearest-neighbor chain algorithm . 10 2 Global–Local Equivalence 13 2.1 Combinatorial optimization problems . 16 2.1.1 Maximum-weight independent set . 22 2.2 Multi-fragment TSP . 25 2.3 Shortest common superstring . 28 2.4 Conclusions . 36 3 Nearest-Neighbor Chain Algorithms 37 3.1 Preliminaries: nearest-neighbor data structures . 37 3.2 Soft nearest-neighbor data structure . 40 3.2.1 Implementation . 41 3.2.2 Choice of parameters . 43 3.2.3 The m-way soft nearest-neighbor data structure . 47 3.2.4 The closest pair problem . 47 3.3 Geometric TSP . 48 ii 3.3.1 Related work .
    [Show full text]
  • Workshop on Probabilistic and Extremal Combinatorics (19W5009)
    Workshop on Probabilistic and Extremal Combinatorics (19w5009) Penny Haxell (University of Waterloo) Michael Krivelevich (Tel Aviv University) Benny Sudakov (ETH Zurich) September 1-6, 2019 1 Overview of the Field Combinatorics, also known as Discrete Mathematics, is a fundamental mathematical discipline which studies the nature of discrete mathematical objects, such as permutations, graphs and set-families. Encompassing some of the most basic and natural mathematical questions, Combinatorics is probably as old as the human ability to count. During the twentieth century and the first two decades of the present century, the field has experienced remarkable growth, transitioning from a collection of mostly isolated results relying on ad hoc arguments, to a major branch of mathematics with its own theory, problems and techniques. Modern Combi- natorics has well-established and fruitful connections to numerous scientific fields, including Number Theory, Discrete Geometry, Statistical Physics, Evolutionary Biology and, most notably, Theoretical Computer Sci- ence, with which it is intimatively intertwined. Combinatorial methods and ideas have been exported to many of the above disciplines, in some cases leading to outstanding results. The workshop focused on two major branches of Combinatorics: Extremal Combinatorics and Proba- bilistic Combinatorics. These two highly-related fields have flourished in recent decades, producing several breakthroughs and many fascinating results, and leading to the creation of a vibrant community comprising both senior and young researchers. We now briefly overview each of these two disciplines. Extremal Combinatorics encompasses several important areas of research, such as Extremal Graph (and Hypergraph) Theory, Extremal Set Theory, and Ramsey Theory. Extremal Graph Theory typically asks to maximize/minimize some graph parameter over all graphs satisfying a given property.
    [Show full text]
  • Hypergraph Matchings and Designs
    P. I. C. M. – 2018 Rio de Janeiro, Vol. 4 (3131–3154) HYPERGRAPH MATCHINGS AND DESIGNS P K Abstract We survey some aspects of the perfect matching problem in hypergraphs, with particular emphasis on structural characterisation of the existence problem in dense hypergraphs and the existence of designs. 1 Introduction Matching theory is a rich and rapidly developing subject that touches on many areas of Mathematics and its applications. Its roots are in the work of Steinitz [1894], Egerváry [1931], Hall [1935] and König [1931] on conditions for matchings in bipartite graphs. After the rise of academic interest in efficient algorithms during the mid 20th century, three cornerstones of matching theory were Kuhn’s ‘Hungarian’ algorithm (Kuhn [1955]) for the Assignment Problem, Edmonds’ algorithm (Edmonds [1965]) for finding a maximum matching in general (not necessarily bipartite) graphs, and the Gale and Shapley [1962] algorithm for Stable Marriages. For an introduction to matching theory in graphs we refer to Lovász and Plummer [2009], and for algorithmic aspects to parts II and III of Schrijver [2003]. There is also a very large literature on matchings in hypergraphs. This article will be mostly concerned with one general direction in this subject, namely to determine condi- tions under which the necessary ‘geometric’ conditions of ‘space’ and ‘divisibility’ are sufficient for the existence of a perfect matching. We will explain these terms and discuss some aspects of this question in the next two sections, but first, for the remainder of this introduction, we will provide some brief pointers to the literature in some other directions.
    [Show full text]